Recurrent Pregnancy Loss and PAI-1: Thrombosis vs. Hypofibrinolysis as the Pathogenic Mechanism?

Well hello again readers! It appears that the 'comment screening' process has been rejuvenated and we are back in business. As feared, however, now I've got to respond to the 50 comments that miraculously appeared in my mailbox yesterday after a week's void. I will do the best I can, but please bear with me while I play catch up between my physical therapy, post-op doctor's appointments, and work!

Below is a great comment from a reader who had the motivation (and fortitude) to tackle my recent series regarding polymorphisms of plasminogen activator inhibitor-1 (PAI-1) and poor pregnancy outcome. I am offering my response to our general readership because it summarizes concisely the major point of enlightenment I had myself during my reasearch on this subject...

Anonymous said...
Dr T.,
Thank you for this very interesting series on PAI-1. From my own personal experience with this defect (4G/4G mutation), I have found that most doctors are still not sure what to make of the experimental data to date. In general, I have received the following comments: 1) I don't think you have a blood clotting problem because you have no history of blood clotting even when you were on birth control pills. 2) I don't test for PAI-1 because regardless of the result I would suggest a person with repetitive miscarriages of your type try Lovenox therapy. 3) I think you do have blood clotting concerns but I don't think this is because of PAI-1. I suspect you have a yet-to-be-discovered variant of Factor V Leiden.

As for the first comment, my only thoughts are that somehow the interaction of PAI-1 and PAI-2 might be responsible for the blood clotting concerns during pregnancy. During one prior miscarriage, it was documented that my Protein S level plummeted in the early weeks of pregnancy, suggesting clotting activity was occurring. I continue to watch for news about PAI-1 and other clotting-related causes of pregnancy loss. I am hopeful that your series will encourage others to look more closely at this interesting gene.Fri Dec 07, 04:20:00 PM 2007


Kenneth F. Trofatter, Jr., MD, PhD said...
To Anonymous Dec 7: Thank you for your thoughtful response. What led to your being tested for PAI-1 in the first place? Was it repetitive early pregnancy losses or other pregnancy-related complications? Anyway, one of the things that dawned on me while writing this series is that the issues related to thrombosis (clotting) and those related to hypofibrinolysis, though related, may actually be SEPARATE issues when it comes to pregnancy complications. What I tried to point out is that even though hypofibrinolysis can lead to thrombosis, it might also lead to other defects (impaired ovulation and/or implantation) that are not related at all to thrombosis! I think this is a subtle point that most providers have missed. Of course when both occur, you've got the double whammy to contend with regard to suboptimal pregnancy outcome.

Incidentally, your doctor may be correct - you could have some other problem that we have not gotten smart enough yet to figure out and, he/she is also correct that an empiric course of therapy is often warranted even if we don't know what we are treating! By the way, protein S levels usually do drop, even during normal pregnancies, although usually this is after the first trimester. You might want to read a post that I wrote earlier this year regarding my first experience (many years ago) with a patient with repetitive miscarriages who only developed a 'lupus anticoagulant' after she conceived. After she got over that 'immunologic hurdle', she did just fine. I have often wondered if activation of the clotting system and the complement system early in pregnancy reflects a suboptimal immune response to pregnancy that can be overcome with time. Could that be your problem as reflected in the rapid drop of protein S activity? Again thank you for reading and best of luck to you!
Dr T

Wed Dec 12, 06:38:00 PM 2007

Labels: PAI-1, recurrent pregnancy loss

Permalink | 9 Comments| Email Post

Post your comment

Plasminogen Activator Inhibitor-1 (PAI-1): Role in Adverse Pregnancy Outcome? - 6 - Treatment and Response Accompany Improved Outcomes

In the last several posts, we have presented support from the literature that links imbalances in the fibrinolytic system, as reflected in increased activity of PAI-1, or a genetic predisposition for the same, with adverse pregnancy outcome, both late and early in pregnancy. With regard to specific mechanisms contributing to recurrent pregnancy loss (RPL) in early pregnancy, we have reviewed evidence to support that aberrations of PAI-1 production could potentially have deleterious effects on ovulation, establishment and maintenance of the corpus luteum (which is essential for ‘progesterone support’ of early pregnancy), and early implantation/placentation of the embryo. In our final post (whew, finally!) on this subject, let’s look at data that would support the premise that down-regulation of PAI-1 production can be accomplished during pregnancy, or in anticipation of pregnancy, in ways that would be safe for both mother and baby, and might improve pregnancy success. At the outset, let me tell you that there is a plethora of information regarding ‘treatment’ of PAI-1 abnormalities under various clinical circumstances, but I would like to highlight only a few articles that pertain directly to pregnancy and RPL.

Bremer and colleagues in 1995 (Am J Obstet Gynecol 1995;172:986-91) performed a small study in which they “assessed the effects of a daily oral dose of 60 to 80 mg of aspirin from 12 weeks gestation until delivery on fibrinolytic variables before and after parturition…in 24 patients, eight receiving low-dose aspirin and 16 controls…The only maternal fibrinolytic variable affected…was plasminogen activator inhibitor activity, which showed a significant reduction before and after parturition of 40% and 70%, respectively, in low-dose aspirin users compared with controls.” None of these patients were reported to have significant pregnancy complications. Since aspirin is not known to have a direct effect on PAI-1 production or activity, it was concluded that the reduction in PAI activity is probably the result of inhibition of platelet reactivity.

In another study published the same year, Gris and colleagues (Thromb Haemost 1995;73:362-7) identified 30 women with a history of unexplained RPL and “an impaired fibrinolytic capacity.” Without identifying the specific reasons for their fibrinolytic imbalances, these women were randomized to begin prior to conception either low-molecular weight heparin (enoxaparin) 20 mg per day or a phenformin-like substance, moroxydine chloride, 1200 mg per day. After one month of treatment, if their fibrinolytic status normalized, therapy was continued for 6 months with the intention to continue treatment if they became pregnant; and, if their fibrinolytic status did not improve after one month, they were switched to the drug they did not receive the first time. The results were actually quite dramatic. With regard to normalization of fibrinolytic status, 20 out of 29 women responded to the first- or second-line enoxaparin treatment whereas did only 1 of 19 treated with moroxydine. Sixteen of 20 enoxaparin responders conceived compared to only 2 of 10 nonresponders (p = 0.002); and, 13 of 16 enoxaparin responder pregnancies resulted in live births compared to none of the 2 nonresponders (p = 0.02). This was compelling evidence that ‘anticoagulation therapy’ at subtherapeutic levels with a heparin compound might improve pregnancy outcome in women identified to have underlying hypofibrinolytic imbalances without even delving into the specific causes of these imbalances.

In 2000, Bick (Clin Appl Thromb Hemost 2000;6:115-25) reviewed the results of anticoagulation therapy in women with histories of RPL and no identifiable chromosomal, hormonal, or anatomical defects. Of the 160 women analyzed, 150 (94%) were found to have coagulation defects, and 38 were found to have more than one of the defects for which they were screened. Their mean age was 33 years and their mean number of miscarriages before referral was three. 149 women were treated preconceptionally with aspirin (81 mg/day) and, immediately following conception, were begun on unfractionated heparin 5000U every 12 hours, both of which were continued until delivery. The remarkable results of this study were that only 2 of the 149 women failed therapy and to have a live birth. This translates to a ‘success rate’ of 98%! In a subsequent report (Bick and Hoppensteadt, Clin Appl Thromb Hemost 2005;11:1-13), among 351 women with RPL who had no other identifiable cause, 322 (92%) were found to have coagulation abnormalities. Those with ‘thrombophilias’ were treated preconceptionally with aspirin (81 mg/day), to which was added following conception, unfractionated heparin (5000U/24 hr) in the first 120 patients, or the low molecular weight heparin, dalteparin (5000U/day), in the next 192 patients. (Patients with MTHFR polymorphisms were also treated with folic acid 5 mg /day and pyridoxine (vitamin B6) 50 mg/day). As the authors reported, “Only 2 of the thrombophilia patients suffered another miscarriage; all others had a normal term delivery” for an overall success rate of 94%.

Also in 2000, Glueck and colleagues (Fertil Steril 2000;74:396-7) presented a case report of a 32 year old woman with amenorrhea and infertility associated with polycystic ovary syndrome (PCOS) who had failure of 7 out of 10 IVF embryo transfers, 1 premature live birth, and two pregnancy losses at 8 and 17 weeks. She was obese, had high fasting serum insulin, androstenedione, and testosterone levels, and was also found to have a modest deficiency in protein S and the 4G4G polymorphism of PAI-1, accompanied by high PAI-1 activity. The combination of the protein S deficiency and the elevated PAI-1 characterized her as having “familial thrombophilia and hypofibrinolysis.” Although not overtly ‘diabetic’, she was begun on metformin (2.55 g/day) and a weight reduction program. Metformin is an oral drug used to treat type 2 diabetes. It improves blood sugar control by various mechanisms, decreasing glucose production by the liver, decreasing absorption of glucose in the gastrointestinal tract, and probably, most importantly, by increasing insulin sensitivity, accompanied by improved peripheral glucose uptake and utilization. Over the course of 4 months, her weight fell from 109 to 91.3 kg (16%), her insulin, androstenedione, and testosterone levels normalized, as did her PAI-1 activity levels.

As a follow-up to this case report, Glueck’s group (Fertil Steril 2001;75:46-52) reported preliminary results from an ongoing pilot study to determine whether metformin could reduce the rate of first trimester pregnancy loss in women with PCOS. They identified 19 women with PCOS who did not have overt diabetes and placed them on metformin (1.5-2.55 g/day) throughout pregnancy. Ten of the women had previously conceived but had miscarried 16 of their 22 pregnancies (73%). “While receiving metformin, these 10 women had 6 normal live births (60%), 1 spontaneous abortion (10%), and 3 normal ongoing pregnancies (30%)” all > 13 weeks. Up to the time of the report, among all 19 women receiving metformin, 9 (47%) had normal term live births, 2 (11%) had normal, but preterm births at 33 and 35 weeks, 6 (32%) had normal ongoing pregnancies beyond 1 weeks, 2 (10.5%) had first trimester miscarriages. No adverse maternal side-effects, nor birth defects were attributed to metformin in this small study. Most importantly, for purposes of our discussion here, “among women who received metformin before conception, reductions in insulin and plasminogen activator inhibitor activity were correlated (r = 0.65; P = .04).” Thus, metformin alone appeared to improve pregnancy outcome in a group of PCOS patients who had either had, or were at increased risk, for early pregnancy loss.

In a subsequent prospective study, Glueck and colleagues (Clin Appl Thromb Hemost 2004;10:323-34) evaluated the efficacy of combined therapy with metformin (1.5 to 2.55 g/day) and enoxaparin (60 mg/day) in women with PCOS and one or more previous early pregnancy losses, thrombophilia, and/or hypofibrinolysis. “Of the 24 women, 23 had 65 previous pregnancies…with 18 live births, 46 spontaneous abortions (71%), and one elective abortion.” Of these 23 women, seven had 3 or more consecutive losses, two had 2 consecutive losses, thirteen had 1 loss, and one woman had a live birth in a pregnancy complicated by HELLP syndrome. Compared to ‘controls’ with no history of adverse pregnancy outcome, the 24 women in this study had a higher frequency of the factor V Leiden mutation (17% vs.2%; P = 0.016), the PAI-1 4G4G polymorphism (46% vs 24%; P = 0.031), higher levels of the PAI-1 gene product and PAI-1 activity (33% vs 8%; P = 0.018), and a higher frequency of elevated factor VIII levels (22% vs 0%; P = 0.037). Of the 23 women who conceived on enoxaparin-metformin to date in the report, they had had 26 pregnancies (28 fetuses), with 20 live births, two ongoing pregnancies > 13 weeks, and 6 spontaneous early losses (21%), 3.4-fold lower than in their previous pregnancies. Again, no adverse maternal or fetal therapy effects were noted.

The articles cited above, have suggested that under various conditions associated with fibrinolytic imbalance and RPL, correction of the imbalance is at least a marker for, if not a direct contributor to, improved first trimester pregnancy success. None of these studies have really confirmed that improvement in pregnancy outcome could be directly correlated with reduction in PAI-1 activity. In the early 1980’s it became recognized that women with PCOS, who ovulated poorly or not at all, would sometimes benefit from partial removal (wedge resection) of their ovaries, or even complete removal of an ovary. This was frequently accompanied by spontaneous ovulation and a decrease in the male hormones (androgens) that can be produced in excess by the ovaries of women with PCOS. Various techniques were employed over the years to reduce the ovarian tissue mass that resulted in the hormonal imbalances accompanying PCOS, but in 1989, Daniell and Miller (Fertil Steril 1989;51:232) described a laparoscopic technique termed ‘ovarian drilling’ in which 4-20 hormone-producing follicles (cysts) on one or both ovaries were pierced and cauterized using laser or electrocautery techniques. This procedure resulted in a dramatic decrease in male hormone levels within days, spontaneous ovulation in 70-90% of women, and a 40-60% probability of pregnancy within a year. Palomba and colleagues (Fertil Steril 2005;84:761-5) performed a comparative study of women with PCOS and elevated PAI-1 levels who underwent ovarian drilling with or without treatment with metformin. Ovarian drilling alone did not reduce PAI-1 activity, whereas metformin administration did. Furthermore they found that a lack of decrease in PAI-1 activity was related to a high risk of miscarriage in those women who conceived following ovarian drilling. These findings suggest that fibrinolytic imbalance, characterized by elevated levels of PAI-1, is an independent risk factor for RPL.

In closing, let me return to one more study by Glueck and colleagues (Metabolism 2006:55:345-52) that we cited in an earlier post. We mentioned previously that results in this study of women with PCOS demonstrated that PAI-1 activity was independently and positively associated with risk for first trimester miscarriage and that “for each 5 IU/mL increment in PAI-1 activity, the risk of being in an adverse first-trimester miscarriage category increased.” What we did not mention before is that prospectively, women in this study were placed on metformin prior to conception and their subsequent pregnancy outcome was assessed and correlated with changes in PAI-1 activity. “From pretreatment to the last preconception visit on glucophage (metformin), in 30 women who subsequently had live births, PAI-1 activity fell 44%, but rose 19% in 23 (also metformin treated) women with first-trimester miscarriage (P = 0.03).” Furthermore, “in the 30 women with live birth pregnancies, median PAI-1 activity fell continuously through the first trimester…, whereas PAI-1 activity was either unchanged or rose in women with first-trimester miscarriage.” Therefore, not only is increased PAI-1 activity an independent risk factor for RPL, but failure of response to therapy, as reflected in lack of normalization of PAI-1 levels also appears to be as well.

At the outset of this series, I dedicated the work to one of our readers (IR) who has had recurrent miscarriages and asked me many months ago about my thoughts on the relationship of PAI-1 activity and RPL. During the course of reviewing the literature I have come to the conclusion that there certainly is a correlation between the two. Perhaps the most compelling evidence resides in the observations that ‘appropriate’ PAI-1 activity appears to be a part of normal ovulation and implantation/placentation in early pregnancy. It is easy to speculate how imbalances at these critical times could interfere with early pregnancy success, regardless of the underlying causes that led to these imbalances. It would appear that efforts to improve a ‘hypofibrinolytic’ state, reflected in women with RPL and increased PAI-1 activity, should be considered as part of any therapeutic regimen. Preconceptional weight reduction, if indicated, treatment with metformin, low-dose aspirin, supplemental folic acid and B-vitamins, ovulation induction with progesterone support, and prophylactic use of heparin or low-molecular weight heparin under these conditions are all options that can be employed and have a wide margin of safety for both mother and baby.

So, IR, I hope this helps. And, I really hope that sometime soon you conceive the baby you are destined to carry. I know you will be a great Mom!
Dr T

Labels: fibrinolysis, PAI-1, recurrent pregnancy loss

Permalink | 14 Comments| Email Post

Post your comment

Plasminogen Activator Inhibitor-1 (PAI-1): Role in Adverse Pregnancy Outcome? - 5 - Implantation and Placentation

Things have been so busy at work the last few weeks that it has been hard to find the time to write here, but I would like to finish up the series we started on plasminogen activator inhibitor-1 (PAI-1) and role in pregnancy outcome and recurrent pregnancy loss (RPL) before moving on to any other major topic. I will admit that the review of the literature on this topic has been somewhat tedious and at times confusing, but it has also been very educational from my own standpoint, and I figure, it’s okay to be a little selfish every once and awhile, especially if a few of our readers might also benefit from the results of the efforts. Anyway, in our last post we reviewed observations that have been made regarding the role of PAI-1 in ovulation and development (and regression) of the corpus luteum and posited possible deleterious effects on these events, under circumstances in which PAI-1 is elevated, that might lead to ovulatory defects and RPL in early pregnancy. Today we are going to turn our attention to observations that have been made regarding the role of PAI-1 in implantation and placentation.

During the first part of the menstrual cycle, prior to ovulation, the hormone that stimulates the growth and the development of the lining (endometrium) of the uterus is estrogen. As we discussed in our last post, following ovulation, the corpus luteum produces progesterone that prepares ('decidualizes') the endometrium to receive the fertilized egg; and, if implantation of the embryo occurs normally, the hormones (mostly hCG) that are produced by the developing placental tissues (trophoblast cells), stimulate the corpus luteum to continue to make progesterone until the placenta gets big enough to take over that function. In 1996, Lockwood and Schatz (J Soc Gynecol Investig 1996;3:159-65) reported that progesterone stimulation (decidualization) of estrogen-primed endometrial cells (estrogen up-regulates endometrial cell receptors for progesterone), both in vitro and in vivo, results in “a marked increase in the expression of tissue factor (TF) and type-1 plasminogen activator inhibitor (PAI-1) and an inhibition of tissue-type and urokinase-type plasminogen activators (tPA and uPA,respectively), matrix metalloproteinases (MMP), and endothelin-1 (ET-1) expression…” Thus, even prior to implantation, the fibrinolytic balance is shifted toward preservation of the extracellular matrix (ECM) structure and prevention of hemorrhage that could be imagined to occur as the embryo implants and the trophoblasts start eating their way into the maternal endometrium. Indeed, progesterone withdrawal shifts the fibrinolytic balance in the opposite direction and this is thought to be one of the factors that permits the shedding/bleeding from the endometrium with menstruation if pregnancy has not occurred successfully.

During the course of normal placentation, some trophoblasts invade the maternal endometrium to anchor the placenta, some migrate through the endometrium and then invade and remodel (open up) the maternal blood vessels (spiral arterioles) that will be the source of blood to the ‘placental bed’ from which the baby will extract oxygen and nutrients (and into which it will transfer its own waste products), and other trophoblasts develop into the frond-like villi that eventually dangle into the placental bed of maternal blood and actually perform these latter functions (transferring oxygen, nutrients, and wastes between the fetal and maternal circulations). Proliferating and invasive trophoblasts produce urokinase-like plasminogen activator (u-PA) and this helps to degrade the ECM and facilitates their migration within the endometrium. One of the primary roles of PAI-1 in normal placentation appears to be in controlling the degree to which trophoblastic cells actually invade the maternal tissues. Graham (Placenta 1997;18:137-43) found that a substance produced by maternal decidual cells, transforming growth factor-beta (TGF-beta), stimulates fetal trophoblasts to make both PAI-1 and the inhibitor of metalloproteinase-1 and also down-regulates trophoblast production of u-PA. This suggests that the trophoblasts, under the direction of maternal cells, may limit their own invasiveness by the secretion of these inhibitors that inactivate u-PA and prevent the degradation of the ECM.

This hypothesis is supported in a study by Floridan and colleagues (Placenta 2000;21:754-62). Beginning with the observation that “trophoblast invasion…in normal intrauterine pregnancies appears to be strictly regulated…whereas tubal and molar pregnancies seem to be characterized by uncontrolled excessive placental invasion,” PAI-1 localization was evaluated in both maternal and fetal tissues under these conditions. In normal pregnancies, PAI-1 was localized, predominantly to trophoblasts that were in closest proximity to maternal tissues: the basal plate of the placenta; the extravillous interstitial trophoblasts comprising the placental anchors in the endometrium; and the trophoblasts that replace maternal vascular endothelial cells as the result of the remodeling of the spiral arterioles. In the basal plate at the deepest layer of placental invasion, PAI-1 (secreted by the trophoblasts?) was noted to be associated with the surface membranes of maternal decidual cells “or confined to the extracellular matrix (ECM) facing the invasive front of anchoring villi.” In contrast, there was a paucity of PAI-1 expression by fetal trophoblasts and maternal cells in both tubal ectopic and molar pregnancies that accompanied uncontrolled trophoblast invasion and damage to maternal tissues.

So, if normal placental invasion of both the decidua and maternal spiral arterioles is at least partly the result of controlled and ‘normal’ expression of PAI-1 activity, we can easily speculate on the potential consequences of excessive PAI-1 production without defining the underlying cause. This could lead to very shallow invasion of the endometrium and inadequate migration to and invasion and remodeling of the spiral arterioles. If severe enough, this could result in early pregnancy loss and if less severe, could result in an abnormal placenta, typical of that seen in preeclampsia, accompanied by increased resistance to fetal and/or maternal perfusion and restriction of fetal growth.

There are many factors that might contribute to the increased expression of PAI-1 under these circumstances. In other posts we discussed the genetic polymorphisms such as 4G/4G that are associated with increased PAI-1 production and activity and this could possibly contribute from both the fetal and maternal sides. Since PAI-1 expression by trophoblasts is at least in part influenced by TGF-beta production by maternal decidual cells, then anything that contributed to increased TGF-beta production by either decidual cells, or cellular components of the maternal immune system, might also result in the fetal trophoblasts producing excessive PAI-1. Indeed, one could imagine that if an abnormal maternal immune response to the pregnancy, either innate or specific, was accompanied by the production of factors that up-regulated PAI-1 expression, this could also contribute to inhibition of trophoblast migration and invasion.

Personally, I believe that aberrations in the maternal immune response are a major cause of abnormalities of placentation that result in pregnancy loss and preeclampsia and there is some evidence indicating that modulation of PAI-1 expression is indeed one mechanism by which these deleterious affects are mediated. Bauer and colleagues (J Clin Endocrinol Metab 2004;89:812-22) reported that the cytokine, tumor necrosis factor (TNF) alpha, inhibits both invasion and migration of trophoblasts in tissue culture experiments. Furthermore, this inhibition is correlated with increased production of PAI-1 and can be reversed by specific antibodies against PAI-1, restoring normal trophoblast migration. In subsequent studies, Renaud and colleagues (Biol Reprod 2005;73:237-43) demonstrated that activated macrophages (components of the innate immune system) also inhibit trophoblast migration. They showed that this inhibition is clearly the result of TNF-alpha production by the macrophages, requiring specific binding of TNF-alpha to the trophoblasts, and that it is accompanied by decreased production of u-PA and increased production of PAI-1 by the trophoblasts themselves.

In our posts to this point, we have built the case that aberrations in the fibrinolytic balance, associated with increased production of PAI-1, accompany adverse pregnancy outcome both late in gestation and in early pregnancy, and have presented several mechanisms by which early pregnancy success might be impaired by increased PAI-1 expression. The questions then remain: can PAI-1 expression be down-regulated in individuals with increased PAI-1 and/or a genetic predisposition for the same and, if so, is this accompanied by improved pregnancy outcome? In our final installment in this series, we will present evidence that it can be and does!

Labels: PAI-1, placental abnormalities; spiral arterioles, recurrent pregnancy loss, RPL

Permalink | 0 Comments| Email Post

Post your comment

Plasminogen Activator Inhibitor-1 (PAI-1): Role in Adverse Pregnancy Outcome? - 4 - Ovulation and Corpus Luteum Development

In our last post on this topic, we concluded that imbalances of the fibrinolytic system are commonly found in women with no other obvious explanations for recurrent early pregnancy loss (RPL) and that one of the more consistent findings in this imbalance is elevated activity of plasminogen activator inhibitor-1 (PAI-1) and a genetic predisposition for the same. Today we are going to begin to review mechanisms by which increased PAI-1 activity could have a deleterious affect on early pregnancy success. In my search through the literature, I first came across several articles that addressed the role of PAI-1 and the fibrinolytic system in placentation. But, I also found several intriguing articles that addressed their roles in ovulation and in the development and regression of the corpus luteum. So let’s start our discussion at that point, since aberrations with ovulation and subsequent development of the corpus luteum might also increase the risk of RPL, and then we will move on to the observations that have been made related to placentation.

First, let me state that this is a very complex subject and in no way do I profess any specialized expertise in this area. Indeed, if any of our readers do have this expertise and want to set me straight or have another interpretation of the scientific literature, please feel free to comment. I am fully aware of my limitations and am here to learn too! Secondly, I am only going to focus on observations related to the fibrinolytic system. I am not going to address the plethora of factors that might affect the actual expression of the various fibrinolytic system components, plasminogen activators or activator inhibitors. So, I ask forgiveness for oversimplification at the outset…

During each normal menstrual cycle, ovarian follicles are ‘recruited’ to produce the ‘egg of the month’; these must undergo development and differentiation and usually one (and sometimes more) becomes the ‘dominant follicle’ and differentiates further. The follicle is comprised of an outer perimeter of extracellular matrix, theca, and granulosa cells that surround the egg and its accompanying (follicular) fluid. At the appropriate time in the cycle, under the influence of luteinizing hormone (LH), the extracellular matrix surrounding the follicle rapidly degrades, the cells at the leading edge of the follicle migrate to and penetrate the surface (capsule) of the ovary, and the egg and its immediate layers of surrounding cells detach from the inner layer of the follicular granulosa cells and is released from the ovary. Once free, the fallopian tube can embrace it, introduce it to friendly sperm along the way, and facilitate the ultimate journey of the fertilized egg to the uterus.

Following ovulation, what’s left of the follicle (the corpus luteum) begins to make the hormone progesterone that helps to prepare (decidualize) the lining of the uterus (the endometrium) to receive the fertilized egg, aiding attachment and implantation of the early embryo. There is a very narrow ‘implantation window.’ With implantation, the fetal cells begin to proliferate and invade the endometrium and they send hormonal messages back to the corpus luteum instructing it to remain healthy and to continue production of progesterone. If conception does not occur, or implantation fails too early, the corpus luteum degenerates, progesterone production drops, and the menstrual cycle starts all over again. So, where does the fibrinolytic system play a role in this sequence of events? Quite frankly, roles have been proposed for every step of this process, but let me focus on just a couple of areas of interest.

Remember from our first post on this subject, plasminogen activators (PA), such as tissue plasminogen activator (t-PA) and urokinase-like plasminogen activator (u-PA), convert plasminogen to its active form, plasmin, which can then direct the breakdown (fibrinolysis) of clots. Well, plasmin actually is in a class of compounds called serine proteases that can help breakdown and rearrange many protein compounds other than fibrin. With regard to the ovary (and other tissues), the extracellular matrix (ECM) is composed of a network of cross-linked protein compounds (mostly collagen) and ‘ground substance.’ In order for the follicle to develop, migrate to the surface of the ovary, and penetrate its surface, this ECM must be loosened up. Plasmin appears to play an important role in the remodeling of the ECM, thereby facilitating follicular cell migration, the actual rupture of the follicle and, perhaps, release of the egg from the follicle as well. PAI-1, which can inhibit both t-PA and u-PA (thereby limiting the production of plasmin), appears to play a key role in the regulation and balance of these events around and within the follicle and probably also plays a role in protecting less developed follicles from destruction as the result of plasmin and other proteolytic enzymes produced at the time of the LH surge.

In various animal studies and studies of tissue samples taken from humans in the periovulatory period, there is certainly evidence to support a role of the fibrinolytic system in the events culminating in ovulation and subsequent development (and regression) of the corpus luteum. Without belaboring this issue, I would simply like to make a few key points by citing a couple of articles on this complicated subject. Tsafriri and Reich (Exp Clin Endocrinol Diabetes 1999;107;1-11) reviewed the data to support the role of the fibrinolytic system in the degradation of the ECM. Following the LH surge, preeovulatory follicles are stimulated to make “a cascade of proteolytic enzymes, including plasminogen activator, plasmin, and matrix metalloproteinases (MMPs). These enzymes bring about the degradation of the perifollicular matrix and, most notably, the decomposition of the meshwork of collagen fibers which provides the strength to follicular wall. Pharmacological blockage of any of these enzymes resulted in the reduction of ovulation rate.” Instead of a “pharmacological blockade”, we could certainly imagine conditions wherein there is increased PAI-1 activity (reducing the activity of these proteolytic enzymes), inhibiting, or simply delaying ,ovulation and, thereby, disturbing the timing related to the ‘implantation window’ of endometrial receptivity and, perhaps, increasing the risk of implantation failure and early pregnancy loss.

Along the same lines, but dealing specifically with regard to the events surrounding follicular rupture and ovulation itself, several observations can be made. In granulosa cells from preovulatory follicles in humans, there appear to be very high concentrations of messenger RNA (mRNA) for both PAI-1 and PAI-2 and low concentrations for PA mRNA and this is reflected in a relative excess of PAI-1 over PA in the follicular fluid surrounding the egg (Jones, et al., J Clin Endocrinol Metab 1989;68:1039-45). In other words, until the egg is ready to hatch, inhibition of the proteolytic system predominates. Studies in the pig (Politis, et al., Biol Med 1990;43:636-42), however, indicate that immediately before ovulation occurs, PAI activity decreases and PA and plasmin activity increase, preceding ovulation. Thus we could imagine, just as we did with ECM degradation, under conditions of excess of PAI activity, ovulation might be prevented or delayed. To add support to this concept, excessive amounts of the pituitary hormone prolactin (PRL) are known to be associated with subfertility. Liu and colleagues (Hum Reprod 1997;12:2748-55) demonstrated in a rat model that PRL delays ovulation and this is accompanied by both a decrease in t-PA production and an increase in PAI-1 production.

Finally, let’s briefly mention the corpus luteum itself. The two primary plasminogen activators, t-PA and u-PA, appear to play divergent roles with regard to the CL. t-PA appears to be involved in the regression of the CL if pregnancy does not occur successfully, and its activity is directly correlated with a decrease in progesterone production. However, an “increase in u-PA mRNA and activity in the early stages of CL development is correlated with an increase in progesterone secretion” (Liu Biol Signals Recept 1999;8:160-77) and the development, growth, and support of the CL (Liu, et al., Endocrinology 2003;144:3611-17) that is essential for the implantation and survival of an early pregnancy. So, again we can use our imaginations. Since PAI-1 can inhibit both t-PA and u-PA, perhaps under circumstances of increased PAI-1 activity as seen in certain women with RPL, the balance shifts from u-PA predominance in the CL to t-PA predominance, causing the CL to regress rather than to grow to support an early pregnancy, thereby leading to pregnancy loss due to inadequate progesterone support. No data on this subject have I found…just a thought!

In our next post on this topic, we will (finally) look at the role of the fibrinolytic system in implantation and placentation….

Labels: corpus luteum, ovulation, PAI-1

Permalink | 2 Comments| Email Post

Post your comment

Plasminogen Activator Inhibitor-1 (PAI-1): Role in Adverse Pregnancy Outcome? - 3 - Association with Recurrent Pregnancy Loss

In our last post, we reviewed several studies that would lend support to the hypothesis that women with elevated levels of plasminogen activator inhibitor-1 (PAI-1), and a genetic predisposition for the same, are at increased risk for adverse pregnancy outcome later in pregnancy. Let me make it quite clear that these studies do not prove that increased PAI-1 causes preeclampsia (or any of the other conditions discussed), but they do suggest that PAI-1 may be a contributor to expression of the disease and particularly the more severe forms of preeclampsia. In today’s post, I would like to begin to build a case with the support of the published scientific literature over the years for an association of increased PAI-1 expression/activity (resulting in decreased fibrinolytic – clot breakdown - activity) with recurrent early pregnancy loss…

In 1993, Gris and colleagues (J Lab Clin Med 1993;122:606-15) evaluated the fibrinolytic system in 116 women who had recurrent early pregnancy loss (RPL) of unknown etiology matched with a group of 90 women who had never had an early miscarriage. Seventy-four of these 116 women with recurrent losses were found to have at least one abnormal test for fibrinolysis compared to none of the control group. A subgroup of 56 women who were shown to have decreased fibrinolytic activity in blood samples taken from veins that were intentionally occluded were selected for further evaluation. Seventeen of these women produced about half the amount of tissue plasminogen activator (t-PA) compared to the controls, 21 had elevated levels of PAI-1 activity, and 16 had both low t-PA and high PAI-1. Other abnormalities that were found among the RPL women were elevated levels of PAI-2 (like that made by the placenta) in nine and decreased urokinase-like plasminogen activator (u-PA) in six. The bottomline is that with these imbalances in the fibrinolytic system, decreased PA activity and/or increased PAI activity, there would be an increased tendency for blood clots to form and not be broken down. These findings led the authors to conclude that “activators and inhibitors of the fibrinolytic system are frequently abnormal in primary habitual aborters” and that “impaired plasmin dependent proteolysis (fibrin clot breakdown) in women might favor recurrent abortion by promoting fibrin deposition in early placental circulation or by limiting trophoblast development.”

Subsequent studies have supported and extended the findings and conclusions above. In 1999, Glueck and colleagues (Metabolism 1999;48:1589-95) found a significant correlation in women with polycystic ovary syndrome (PCOS) between elevated levels of PAI-1, early pregnancy loss, and no live births. These authors concluded it “is a predominant independent significant positive reversible risk factor for miscarriage in women with PCOS.” In 2003, Dossenabch-Glaninger, et al. (Clin Chem 2003;49:1081-6) evaluated 49 women with a history of two consecutive, or 3 to 6 nonconsecutive, early pregnancy losses compared to 48 women without a history of pregnancy loss for several genetic variants of the coagulation system. They found that homozygosity for PAI-1 or the factor XIII 34 Leu polymorphisms or compound heterozygous status (both of these polymorphisms in the same individual) of these same mutations significantly increased the risk for early pregnancy loss (OR = 2.4; 95% CI, 1.1-5.5).

In our last post, we mentioned that PAI-1 produced by vascular endothelial cells is induced by angiotensin II which is generated by the action of angiotensin I converting enzyme (ACE) and that autoantibodies directed against the angiotensin II type 1 receptor (AT1) found in preeclamptic women was associated with associated with increased production of PAI-1 (Xia, et al., J Soc Gynecol Invest 2003;10:82-93). Along the same lines, Buchholz and colleagues (Hum Reprod 2003;18:2473-7) studied the ACE deletion(D)/insertion(I) and the PAI-1 4G/5G polymorphisms in women with RPL, both of which are associated with increased ACE and PAI-1 expression, respectively. Comparing 184 women with a history of two or more consecutive spontaneous abortions with 127 women who had term pregnancies and no early losses, they found that homozygosity for the D allele of the ACE gene (D/D) was significantly correlated with RPL and the presence of the PAI-1 4G/4G homozygous state further increased PAI-1 levels and risk for early pregnancy loss. As a consequence of these findings, the authors recommended “the incorporation of these two polymorphisms into the spectrum of thrombophilic mutations which should be analyzed in individuals with recurrent spontaneous miscarriages.”

In more recent studies, Glueck and colleagues (Metabolism 2005;54:1345-9) reported that even among women who had had live births, if they had also had a spontaneous abortion, they were at greater risk than women who had never lost a pregnancy for having elevated levels of PAI-1 (33% vs 18%) and for the presence of other aberrations of the coagulation cascade: presence of factor V Leiden homozygosity (15.2% vs 1.6%) and elevated levels of factor VIII (31% vs 18%). These findings carried over to similar observations in women with PCOS (Glueck, et al., Metabolism 2006;55:345-52). In this study they assessed the association of PAI-1 levels in 430 women with PCOS who were divided into the following groups: 1) women who had live births only (n = 208); 2) women who had one or more live births and one or more first trimester losses (n = 111); 3) women who had only had first trimester miscarriages (n = 71). They found that “PAI-1 activity was positively associated with first-trimester miscarriage (p = 0.004)” … and “for each 5 IU/mL increment in PAI-1 activity, the risk being in an adverse first-trimester miscarriage …increased (OR, 1.12; 95% CI, 1.04-1.20).” In the same study, they also evaluated the association of the PAI-1 4G polymorphism in 921 women with PCOS compared to 126 normal females and again demonstrated (although the difference was not as dramatic in these women with a more heterogeneous obstetrical history – 78% in the PCOS group compared to 69% in controls) the 4G allele “is more common in women with PCOS than in normal women and, in concert with obesity, hyperinsulinemia, and hypertriglyceridemia, contributes to treatable, hypofibrinolytic, miscarriage-promoting, high PAI-1 activity.”

Coulam and colleagues (Reprod Biomed Online 2006;12:322-7) compared the prevalence of ten thrombophilic gene mutations in 42 women with a history of recurrent implantation failure after IVF embryo transfer with 20 fertile women. They found that the women with implantation failure had a significantly higher prevalence of PAI-1 4G/5G polymorphisms than controls (P = 0.007). Although they found no significant differences in the prevalence of any other single gene mutation, they did find “the prevalence of total gene mutations among patients with implantation failure was significantly higher than among controls. More than 3 gene mutations among the 10 genes studied were observed in 74% of women with implantation failure” compared to 20% of controls (P = 0.0004). They “concluded that inherited thrombophilias are associated with implantation failure” and this highly significant “association is manifest by totatl number of mutations as well as with PAI-1 mutations.”

Using the same approach, Coulam’s group also reported (Am J Reprod Immunol 2006;55:360-5) a comparison between 150 women with two or more recurrent pregnancy losses and 20 fertile women with no history of pregnancy losses. In this study they also found that there were “no differences in the frequency of specific gene mutations…however, the prevalence of homozygous mutations (59% vs 10%) and total gene mutations among patients with recurrent miscarriage was significantly higher than among controls.” As in their previous report, more than 3 mutations among the 10 genes studied were observed in a significantly higher percentage of women with recurrent miscarriage than controls (68% vs 21%). Subsequently, they reported on 550 women with RPL and among the polymorphisms they investigated they found that PAI-1 4G/5G (P = 0.009), factor XIII V34L (P < 0.0001), and homozygous MTHFR C677T (P < 0.0001) correlated significantly with RPL compared to controls.

To summarize today’s post, despite the heterogeneity of findings detailed in the studies referenced above, there seems to be a common thread: recurrent pregnancy loss, in the absence of other explanations (e.g., chromosomal abnormalities, uterine anomalies, chronic maternal disease) is frequently accompanied by imbalances in the fibrinolytic system. These imbalances are uniformly those that lead, theoretically, to decreased fibrinolysis and may include genetic defects of the coagulation system and are frequently accompanied by conditions that lead to increased production of PAI-1. The mechanism by which these factors might actually contribute to an increased risk for recurrent early pregnancy loss will be discussed in the next post…

Labels: fibrinolysis, PAI-1, recurrent pregnancy loss

Permalink | 42 Comments| Email Post

Post your comment

Plasminogen Activator Inhibitor-1 (PAI-1): Role in Adverse Pregnancy Outcome? - 2 - Late Pregnancy Complications

In our last post, we discussed the role of plasminogen activator inhibitor-1 (PAI-1) in helping to maintain the balance between the clotting and fibrinolytic (clot-dissolving) sides of the coagulation system. The primary function of PAI-1 is to inhibit plasminogen activators (t-PA and u-PA) from converting plasminogen to plasmin which is responsible for initiating fibrinolysis. The premise is that if there is too much PAI-1 activity, clots will tend to hang around longer and if there is too little, the individual would be at increased risk for bleeding problems. Before we can address possible roles of abnormalities of PAI-1 production and activity in adverse pregnancy outcome and recurrent pregnancy loss (RPL), it would be helpful to understand changes that might occur in these parameters during normal pregnancy.

Kruithof and colleagues (Blood 1987;69:460-6) reported that both plasminogen activators (t-PA and u-PA) and plasminogen activator inhibitors increased during pregnancy. t-PA and u-PA increased 50% and 200%, respectively, throughout normal pregnancy. They also found that PAI-1, produced predominantly by endothelial cells lining blood vessels, increased nearly 10-fold by term over that found in nonpregnant women and a second plasminogen activator inhibitor, PAI-2, not found in nonpregnant women, but produced by the placenta, was present in very high concentrations by term. The increase in both activators and inhibitors appeared to maintain the balance between the clotting and fibrinolytic systems during normal pregnancy because no changes in plasminogen or the overall fibrinolytic activity were found. Within “three to five days after delivery most parameters of the fibrinolytic system were normal again.”

In 1989, Estelles and colleagues (Blood 1989;74:1332-8) reported that women with severe preeclampsia in third trimester had significantly higher levels of PAI-1 than nonhypertensive women. Interestingly, PAI-2 levels were significantly lower in the preeclamptic women and a positive correlation between birth weight and PAI-2 levels was found (in other words, the higher the PAI-2, the greater the birth weight); and birth weight was inversely correlated with PAI-1 levels (higher the PAI-1 activity, the lower the birth weight). The presumption is that the lower PAI-2 levels correlated with a decreased placental mass or function in preeclamptic women. Regardless, the high levels of PAI activity in severe preeclampsia appear to be solely related to the increased activity of PAI-1. And, as many of our readers are aware, this might account in part for the coagulation abnormalities frequently accompanying the more severe forms of preeclampsia.

Unfortunately, these observations late in pregnancy don’t really tell us whether elevated levels of PAI-1 in preeclampsia are a cause, an effect, a response, or a contributor to the disease process itself. Based on several observations by other investigators, and the putative role of PAI-1 in placentation early in pregnancy (which we will eventually get to here), perhaps it is all the above. There does appear to be a genetic predisposition/association with abnormalities in PAI-1 production and later pregnancy complications. Yamada and colleagues (J Hum Genet 2000;45:138-41) evaluated the association between preeclampsia and deletion/insertion polymorphisms (4G or 5G) in the promoter of the PAI-1 gene. The 4G/5G polymorphism was assessed in 115 women with preeclampsia, 210 normotensive pregnant women and 298 nonpregnant controls. The frequency of the 4G allele (which results in increased production of PAI-1) and of 4G/4G homozygosity was significantly higher in the preeclamptic women than either the normal pregnant or nonpregnant controls, suggesting that the presence of 4G is one risk factor for preeclampsia and perhaps more severe manifestations of the disease.

Along the same lines, Glueck, et al. (Metabolism 2000;49:845-52) evaluated complications in 133 women with at least one pregnancy, and found a significant association of the 4G/4G PAI-1 polymorphism with prematurity, intrauterine growth restriction (IUGR), and “total complications of pregnancy” that was independent of the presence of other genetic thrombophilias (factor V Leiden, MTHFR C677T, and prothrombin G20210A mutations). In a subsequent study (Glueck, et al., Obstet Gynecol 2001;97:44-8), they reaffirmed the presence of the 4G/4G genotype as a risk factor for IUGR and extended their findings to include associations with severe preeclampsia, placental abruption, and stillbirth. They also reported that “the hypofibrinolytic 4G/4G mutation of the PAI-1 gene…is frequently associated with the thrombophilic factor V Leiden mutation” which would further increase the risk of problems related to clotting.

Over the years, PAI-1 made by vascular endothelial cells was found to be induced by angiotensin II which is produced by the action of the angiotensin I-converting enzyme (ACE). In a fascinating paper published in 2003, Xia and colleagues (J Soc Gynecol Invest 2003;10:82-93) reported that 18 of 20 women with severe preeclampsia were found to have IgG antibodies to the angiotensin II type 1 (AT1) receptor. None of 18 normotensive pregnant women had these autoantibodies. They also found that the serum from the same 18 of 20 women with these AT1 receptor autoantibodies stimulated PAI-1 secretion by trophoblasts (placental cells) in culture. Activation of the trophoblast AT1 receptors was also correlated with decreased trophoblast migration and invasion in tissue culture models and this, too, was directly correlated with PAI-1 production. We will return to this point in our subsequent discussion of the role of PAI-1 in recurrent early pregnancy loss. Bobst and colleagues (Am J Hypertens 2005;18:330-6) further reported that AT1 receptor autoantibodies found in preeclamptic patients stimulated PAI-1 (and the cytokine IL-6) production by human kidney (mesangial) cells in culture. Reversible ‘damage’ to the kidney is one of the events which characterize preeclampsia and the more severe the kidney impairment, generally, the more severe the preeclampsia with regard to hypertension and decreased urine production...(more to follow!)...

Labels: IUGR, PAI-1, preeclampsia, thrombophilias

Permalink | 12 Comments| Email Post

Post your comment